Yuxuan Zhang
I’m a research intern at Hao AI Lab, advised by Prof. Hao Zhang and Prof. Haojian Jin.
My research focuses on building useful AI agents and efficient LLM reinforcement learning systems.
Recent News
Feb 2025 Incoming to Google as a Software Engineering Intern
Jan 2025 LMGame-Bench: How Good are LLMs at Playing Games? is accepted to ICLR 2025
Projects
1M views
AI agent playing Ace Attorney game with 1 million views
Publications
LMGame-Bench: How Good are LLMs at Playing Games?
ICLR 2026
Introduces **LMGame-Bench**, a unified Gym‑style benchmark that tests LLM agents across platformer, puzzle, and narrative games—addressing vision brittleness, prompt variance, and data contamination.
Paper Code
ICLR 2026
Introduces **LMGame-Bench**, a unified Gym‑style benchmark that tests LLM agents across platformer, puzzle, and narrative games—addressing vision brittleness, prompt variance, and data contamination.
Paper Code
General Modular Harness for LLM Agents in Multi‑Turn Gaming Environments
ICML MAS Workshop 2025
Introduces a perception–memory–reasoning harness that consistently boosts LLM/VLM gameplay across classic and modern game suites, uncovering module‑specific performance patterns.
Paper
ICML MAS Workshop 2025
Introduces a perception–memory–reasoning harness that consistently boosts LLM/VLM gameplay across classic and modern game suites, uncovering module‑specific performance patterns.
Paper
Experience
Education
M.S. in Computer Science
University of California, San Diego
University of California, San Diego
B.S. in Data Science
University of California, San Diego
Summa cum laude (top 2%)
University of California, San Diego
Summa cum laude (top 2%)
Research Experience
Research Intern
Hao AI Lab, UC San Diego
Hao AI Lab, UC San Diego
Research Intern
Scripps Research, La Jolla, CA
Scripps Research, La Jolla, CA
Teaching Experience
Teaching Assistant
DSC204A: Scalable Data Systems, UC San Diego
DSC204A: Scalable Data Systems, UC San Diego